Databricks used its Data + AI Summit on June 16 to introduce CustomerLake, a new agentic customer data platform built natively inside the Databricks Lakehouse. The product brings core CDP functions — Customer 360, identity resolution, audience building, campaign automation, activation, and personalization — directly into the environment where enterprises already store customer data, AI models, and governance controls, with the explicit goal of letting teams act on customer data without moving or duplicating it.
The pitch to marketing and data leaders is architectural. Databricks argues that most enterprises struggle with customer data not because of strategy but because their CDP sits outside their core data and AI platform, creating yet another system to integrate, govern, and reconcile. CustomerLake’s premise is that agents need governed access to identity, predictive models, business logic, activation endpoints, and real-time performance signals all in one place — and that the lakehouse is where that context already lives.
Two agent types at the core
CustomerLake is organized around two agentic capabilities. Profile Agents turn raw customer data into business-ready Customer 360 profiles, preparing data, flagging quality issues, and supporting third-party enrichment to unify fragmented records into golden profiles. Underpinning this is what Databricks calls Agentic Identity Resolution (AIR), an approach that combines deterministic, probabilistic, and agentic workflows and allows teams to bring their own identity rules, models, and enrichment partners.
Campaign Agents handle the activation side, using governed customer context to build audiences, recommend next-best actions, activate across channels, and continuously optimize against business goals. Databricks frames this shift as moving from static, one-off campaigns to what it terms “infinity campaigns” — continuous, agent-driven engagement loops that analyze signals, decide on the next action, and act across channels in real time.
Notably, the company positions humans as still owning strategy, goals, and guardrails, with agents scaling execution rather than replacing oversight. The traditional campaign workflow — define an objective, request data, build a segment, validate, launch, measure, repeat — collapses into defining a goal (such as growing loyalty enrollment or reactivating lapsed customers) and letting Campaign Agents handle identification, eligibility constraints, offer and channel selection, activation, and optimization.
Interoperability and a consumption-based model
CustomerLake is governed by Unity Catalog and uses Lakehouse Federation to access customer data wherever it sits, including Snowflake, Google BigQuery, cloud object storage, and operational databases. Through native integrations and Reverse ETL, it provides bi-directional pipelines into the martech and adtech tools enterprises already use.
The launch partner roster is broad, spanning identity, activation, measurement, and customer experience: Adobe, Meta, Braze, Acxiom, Epsilon, The Trade Desk, LiveRamp, Iterable, Bloomreach, Snapchat, Magnite, TransUnion, Adstra, Twilio, Integral Ad Science, and Unity. Services partners include Accenture, Deloitte, Lovelytics, Slalom, and Stitch. Databricks also notes a “value-aligned consumption model” pitched as a more flexible alternative to traditional software licensing — a clear signal it is targeting enterprises looking to consolidate martech spend.
Early adopters and availability
Several global brands are cited as early users. HP’s Kumar Ram, Global Head of Marketing Technology and AI Enablement, framed the appeal around governed customer context rather than another copy of the data. Circle K’s Jay Malepati described building audiences natively in Databricks, activating them in Adobe, and measuring impact without relocating its data lake. Getnet by Santander’s Chief Data and AI Officer Ainhoa Alonso positioned it as a way to strengthen its CRM and merchant relationships at scale.
CustomerLake is currently available in Private Preview. Co-founder and CEO Ali Ghodsi positioned the launch as a replacement for legacy marketing software, arguing that when customer data, AI models, and agents share one governed platform, marketing shifts from discrete campaigns to a continuous analyze-decide-act loop — and increasingly addresses a new kind of customer: the AI agents that consumers themselves are starting to use to browse and buy.
Why it matters for marketing leaders
The announcement lands squarely in the convergence of the CDP and the data platform, a trend that has been building as warehouse-native and “composable” CDP approaches gain traction. For marketing and CX leaders, CustomerLake raises a familiar build-versus-buy question in a new form: whether the customer data platform should remain a discrete application or collapse into the governed data foundation the rest of the enterprise already runs on. With Adobe, Twilio, Braze, and other established martech players appearing as partners rather than competitors, the practical question for buyers will be how CustomerLake coexists with — or eventually displaces — incumbent CDP and activation investments.
